explaingit

googleapis/mcp-toolbox

★ 15,197GoAudience · developerComplexity · 3/5LicenseSetup · moderate

TLDR

MCP Toolbox is an open-source server that gives AI assistants like Claude or Gemini standard tools to query enterprise databases, from PostgreSQL to BigQuery, without writing custom integration code each time.

Mindmap

mindmap
  root((mcp-toolbox))
    What it does
      MCP server for DBs
      AI talks to databases
      No custom glue code
    Prebuilt Tools
      list_tables
      execute_sql
      Semantic search
    Supported Databases
      PostgreSQL MySQL
      BigQuery Spanner
      MongoDB Redis
    Client SDKs
      Python JavaScript
      Go Java
    Integrations
      LangChain LlamaIndex
      Claude Code Gemini
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Things people build with this

USE CASE 1

Connect Claude Code or Gemini CLI to your PostgreSQL database so you can explore tables and run queries in plain English.

USE CASE 2

Define custom AI agent tools in a YAML file that let an LLM run safe, structured queries against BigQuery or Cloud SQL without backend code.

USE CASE 3

Build a production AI agent with LangChain or LlamaIndex that has authenticated, connection-pooled access to multiple enterprise databases.

USE CASE 4

Prototype a natural-language-to-SQL interface over any supported database using the prebuilt list_tables and execute_sql tools.

Tech stack

GoPythonTypeScriptJavaScriptJavaYAML

Getting it running

Difficulty · moderate Time to first run · 30min

Requires a running database and an MCP-compatible AI client, Google Cloud databases need IAM credentials configured before connecting.

Use freely for any purpose, including commercial use, as long as you include the Apache 2.0 license notice.

In plain English

MCP Toolbox for Databases is an open-source server that lets AI assistants and AI agents talk to enterprise databases. MCP stands for Model Context Protocol, a way of giving AI clients a standard set of tools they can call on. This project plays that role for databases: instead of writing custom glue every time you want an AI to read your tables or run a query, you point it at the Toolbox. It is described as serving a dual purpose. The build-time side is a ready-to-use MCP server with prebuilt generic tools such as list_tables and execute_sql, so you can connect an MCP-compatible client like Gemini CLI, Google Antigravity, Claude Code, or Codex and immediately explore a database in plain English. The run-time side is a framework for building your own production-grade tools, using a tools.yaml file where you declare sources (the databases you want to expose) and tools (the actions an agent is allowed to take, such as structured queries, semantic search, or natural-language-to-SQL). It comes with connection pooling, integrated authentication, and observability through OpenTelemetry. Client SDKs are available for Python, JavaScript/TypeScript, Go, and Java, with integrations documented for the Agent Development Kit, LangChain, and LlamaIndex. Supported databases include Google Cloud's AlloyDB, BigQuery, Cloud SQL, Spanner, Firestore, and Knowledge Catalog, plus PostgreSQL, MySQL, SQL Server, Oracle, MongoDB, Redis, Elasticsearch, CockroachDB, ClickHouse, Couchbase, Neo4j, Snowflake, and Trino. The repository was previously named genai-toolbox. It is written in Go and licensed Apache 2.0. The full README is longer than what was provided.

Copy-paste prompts

Prompt 1
Set up MCP Toolbox to connect Claude Code to my PostgreSQL database. Write the tools.yaml config and tell me what MCP client settings to add in Claude Code.
Prompt 2
How do I define a custom tool in MCP Toolbox tools.yaml that runs a parameterized SQL query against BigQuery and exposes it to an AI agent?
Prompt 3
I want to use MCP Toolbox with LangChain in Python to let an AI agent query my MySQL database. Show me the full setup code using the Python SDK.
Prompt 4
How does MCP Toolbox handle authentication when connecting to Cloud SQL? What do I need to configure for IAM-based access?
Prompt 5
I want to add semantic search on top of my database using MCP Toolbox. Which databases support it and how do I define that tool in the YAML?
Open on GitHub → Explain another repo

← googleapis on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.